Run predictions from state-of-the-art machine learning models in Tensorflow & PyTorch right from your browser. Use intuitive GUIs, no preprocessing or coding required! Inference time may run up to 15 seconds for predictions. For quicker models or API access, contact us.
This Convolution Neural Network was trained on Google's QuickDraw dataset with 345 classes. Try it by drawing a lightbulb, radio, or anything you can think of!.
Upload a picture of yourself or another person, and the Openpose algorithm will identify your face, arms, and legs. Handles multiple people as well!.
Colorize Photos (DeOldify)
DeOldify colorizes old black and white images and photos. Try it out on an old image you want restored! Note: It takes about 30-40s to process the image..
Identifying Skin Cancer
Predicts whether an image of skin is cancerous or not. This model is EXPERIMENTAL and should only be used for research purposes. Please see a doctor for any diagnostic reasons..
How well can this model predict your emotions? Take a picture with your webcam, and it will guess if you are: happy, sad, angry, disgusted, scared, surprised, or neutral..
Draw a number 0 through 9 on the sketchpad, and click submit to see the model's predictions. Model trained on the MNIST dataset..
Surviving the Titanic
What is the probability that a passenger on the Titanic would survive the famous wreck? It depends on their demographics as this live interface demonstrates..
MobileNet vs. InceptionNet
Compare 2 state-of-the-art machine learning models that classify images into one of 1,000 categories: MobileNet (top), a lightweight model that has an accuracy of 0.704, vs. InceptionNet(bottom), a heavier model that has an accuracy of 0.
This model, RNA-GPS, is trained on APEX-seq data, which measures RNA localization within human HEK293T cells. The model predicts localization of RNA transcripts to eight different sub-cellular compartments.
Hotdog or Not?
Based off of Jian Yang's infamous model from the TV show Silicon Valley, this model has one job, just like its name suggests..
Painting Generator (StyleGAN)
This GAN model, trained on WikiArt images, generates a never-seen-before painting based on an artist, style, and genre!.
GPT-2 is a large transformer-based language model with 1.5 billion parameters, trained on a dataset of 8 million web pages. GPT-2 is trained with a simple objective: predict the next word, given all of the previous words within some text.
Left Ventricle Segmentation
This semantic segmentation model identifies the left ventricle in echocardiogram videos. Accurate evaluation of the motion and size of the left ventricle is crucial for the assessment of cardiac function and ejection fraction.
If you'd like to see a specific model uploaded here, let us know. 😊